Original Research Fat Suppression With Spectrally Selective Inversion vs. High Spectral and Spatial Resolution MRI of Breast Lesions: Qualitative and Quantitative Comparisons Xiaobing Fan, PhD, 1 Hiroyuki Abe, MD, 1 Milica Medved, PhD, 1 Sean Foxley, BS, 1 Sanaz Arkani, MS, 1 Marta A. Zamora, BA, 1 Olufunmilayo I. Olopade, MD, 2 Gillian M. Newstead, MD, 1 and Gregory S. Karczmar, PhD 1 * Purpose: To compare conventional fat-suppressed MR im- ages of the breast to images derived from high spectral and spatial resolution MR data. Image quality and the level of fat suppression are compared qualitatively and quantita- tively. Materials and Methods: Women with suspicious breast lesions found on X-ray mammography were imaged on 1.5 Tesla GE SIGNA scanners. High spectral and spatial reso- lution (HiSS) data were acquired using echo-planar spec- troscopic imaging. Images with intensity proportional to the water signal peak height in each voxel were synthesized. Conventional fat-suppressed images were acquired using a frequency selective inversion method. The experimental (HiSS) and conventional images were compared by experi- enced radiologists to evaluate the quality of fat suppres- sion. In addition, fat suppression and image quality were evaluated quantitatively. Results: Fat suppression, tumor edge delineation, lesion conspicuity, and image texture were improved in the peak height images derived from HiSS data. Conclusion: The results demonstrate that the water peak height images obtained from HiSS data potentially could improve the quality of fat suppression, detection and diag- nosis of breast cancer. HiSS allowed detection of lesions and evaluation of lesion morphology prior to contrast media injection. Key Words: echo-planar spectroscopic imaging; breast le- sion; high spectral and spatial resolution imaging; fat sup- pression; fat-saturation J. Magn. Reson. Imaging 2006;24:1311–1315. © 2006 Wiley-Liss, Inc. EARLY AND ACCURATE DIAGNOSIS of breast cancer greatly improves survival and reduces morbidity. MRI has the potential to greatly improve cancer treatment. While MRI clearly increases sensitivity to breast cancer, improvements in specificity are essential before MRI becomes a widely used clinical tool. In addition, further increases in sensitivity are desirable so that very early cancers, such as ductal carcinoma in situ (DCIS), can be detected more reliably (1). To accurately assess the morphology of breast lesions and to delineate the lesion boundaries, high quality fat suppression images are required. Conventional imaging does not make use of spectral information and represents each spatial location or voxel using a single complex number. To improve breast imaging we introduced the high spectral and spatial resolution (HiSS) approach (2,3). HiSS data can be acquired rapidly using echo-planar spectroscopic imaging (EPSI) to provide detailed water and fat spectra associated with each image voxel at the spatial resolu- tion of conventional imaging or greater. HiSS data are analyzed to produce images representing various as- pects of the water or fat resonances. These images in- clude water peak height images, pure T 2 * images, pure T 1 images (4), and Fourier component images (5). Previous work has compared HiSS water peak height images to conventional fat-saturated images, and sug- gested that HiSS images offer significant advantages (4,6). However, because HiSS was incorporated into the standard clinical protocol, the fat-saturated images were not acquired with the same spatial parameters (spatial resolution and slice locations and thickness) as the HiSS water peak height images (4,6), and the fat- saturated images were acquired after contrast injection 1 Department of Radiology, University of Chicago, Chicago, Illinois, USA. 2 Section of Hematology and Oncology, University of Chicago, Chicago, Illinois, USA. Contract grant sponsor: Segal Foundation; Contract grant sponsor: National Institute of Biomedical Imaging and Bioengineering (NIBIB); Contract grant number: RO1 EB003108-01. Contract grant sponsor: National Cancer Institute (NCI); Contract grant number: RO1CA78803. Contract grant sponsor: Army Breast Cancer Research Program; Con- tract grant number: DAMD 17-02-1-0033. *Address reprint requests to: G.K., Ph.D., Professor, Department of Radiology, MC2026, University of Chicago, 5841 S. Maryland Ave., Chicago, IL, USA 60637. E-mail: gskarczm@uchicago.edu Received December 9, 2005; Accepted July 25, 2006. DOI 10.1002/jmri.20732 Published online 9 November 2006 in Wiley InterScience (www. interscience.wiley.com). JOURNAL OF MAGNETIC RESONANCE IMAGING 24:1311–1315 (2006) © 2006 Wiley-Liss, Inc. 1311